As of September 2026: Indian IT still prints cash, but five-year price returns and a shrinking Nifty weight show the market has stopped treating the leaders as automatic compounders.
Published 2026-09-06 · Updated 2026-09-10 · 10 min · By Abhishek · IT · Markets
Indian IT still prints cash. The market no longer pays for it like a growth story.
Clients are still signing contracts. Margins at the leaders are high. Free cash flow is intact. What broke is the old bargain: mid-teens dollar growth, a rising headcount pyramid, and a premium multiple. Growth is now low single digits. Headcount at the top firms is falling. The businesses look solid. The stocks have not.
Two facts make that split concrete. IT’s weight inside Nifty 50 has shrunk to a multi-decade low. And over five years, three of the five leaders have delivered negative price returns. Those are the charts that matter.
Who still leads
India’s broader tech industry (IT services, BPM, engineering R&D, products, hardware) is around $315 billion in FY26, up about 6% (Nasscom). AI-related work is roughly
0–12 billion: real, but still a thin slice.
The listed ranking by scale has not changed:
TCS (~$30 bn) - first annual constant-currency decline since listing; still the scale and margin leader.
Wipro (~
0.5 bn IT services) - flat to slightly down; multi-year slog.
Tech Mahindra - smaller of the five; low single-digit growth; still the junior sibling.
The top four together did about $75 billion. FY27 guidance from the large names sits in a 1–4% band. That is a holding year, not a rebound year.
Deal-rich, job-light
Order books did not collapse. TCS full-year TCV was about $40.7 billion. Infosys large-deal TCV about
4.9 billion. HCLTech new-deal TCV about $9.3 billion. Many new contracts carry an AI or automation clause.
Pricing is the other side. Generative AI makes legacy maintenance cheaper; clients ask for that saving. Research notes often put 2–3% annual deflation on traditional work even as AI work is added. Revenue can look dull while “AI revenue” slides look busy.
Headcount is the visible proof. Combined net headcount at the five fell in FY26. TCS took the sharpest cut (on the order of 23,000–25,000). Freshers are still hired, but the pyramid is no longer the growth engine. Revenue per employee has to do the work that hiring used to do.
Two more facts:
Margins held. TCS moved back into the mid-20s. Infosys near 20–21%. Wipro and HCLTech mid-to-high teens. Cash generation did not break.
GCCs matter. India hosts well over 1,700 global capability centres. Build-operate-transfer work for those centres is still a structural edge for the large Indian firms.
Smaller inside the index
Underperformance does not stay on a sector chart. It changes who runs the benchmark.
IT has gone from a market-defining slab of Nifty 50 to a mid-weight. In two years it lost roughly 6–7 percentage points of index weight; from its old peak, more than half. Passive money follows those weights. At points this cycle, HDFC Bank alone has weighed more than the top five IT firms combined.
Even a good quarter at TCS or Infosys no longer moves Nifty 50 the way IT did in 2009, 2020 or 2021. Leadership has shifted to financials, energy and domestic cyclicals. IT can recover as a trade. It is no longer the default engine of Nifty returns.
Five years, five leaders
The last five years were supposed to be the easy part of the Indian IT story. They were not.
This was not a gentle grind lower. It was a boom, then a long give-back. 2021 still carried the pandemic digital-spend hangover: expensive and loved. 2023–24 offered relief rallies. 2025 and the first half of 2026 took most of that back as growth slowed to low single digits and investors began pricing AI as a threat to billable hours, not only as a new service line.
So the three oldest compounders (TCS, Infosys, Wipro) sit about 34% to 46% below where they were five years ago on price. TCS, still the best business of the lot on margins and cash, is among the worst stocks of the lot. Quality of earnings and quality of return have diverged.
HCLTech and Tech Mahindra held up mainly because the starting valuation was less heroic, and Tech Mahindra never ran as far in the first place. “Held up” means roughly +10% in five years: about 2% a year. That is survival relative to the other three, not a compounding story.
For context, Nifty 50 over long stretches still compounds in the low-to-mid teens in decent cycles. Treating “buy the IT majors and wait” as a five-year rule did not do that for TCS, Infosys or Wipro on price alone.
What is still hard
Discretionary spend in the US and Europe is still tight. BFSI, retail and manufacturing are buying cost-out and productivity, not greenfield transformation.
Vendor consolidation helps the mega-deal winner and starves everyone else.
AI eats legacy billing. New AI revenue and old price cuts can cancel each other on the top line for a while.
Hiring mix has changed: fewer mass freshers, more specialists. Better for margins later; painful for the old volume model now.
Geopolitics and visas add noise. They are not the main plot.
FY27 is fairly described as consolidation and capability-building. The test is simple: can AI-native growth outrun deflation on legacy contracts? Until revenue per employee and outcome-based pricing show up clearly in the P&L, the market will stay sceptical.
Looking ahead
Near term (FY27–FY28). Low-single-digit growth at TCS, Infosys and Wipro; HCLTech and Tech Mahindra more variable. Margins defended by automation, utilisation and a weaker rupee, not by pricing power. Stocks can rally on a raised guide or a beat of a low bar, and fall again on one weak global peer print. Until growth turns, this remains a trading sector.
If execution holds from FY29. AI mix has to become material *and* net new, not a relabel of old work at a lower price. Outcome-based deals, platforms and GCC work have to replace hours sold.
Among the five. TCS remains the default quality name. Infosys is the other core franchise if large-deal wins stay high. HCLTech has a slightly different services-plus-products mix. Wipro is still the turnaround. Tech Mahindra is the higher-beta junior. The revenue ranking has not changed. The valuation and the index weight have.
Conclusion
Put the two charts next to the P&L and the picture is consistent.
The weight chart says the market’s verdict is already written into the benchmark. IT has slid from about 14% of Nifty 50 at end-2024 to under 8% by mid-2026, and from about 21% at its old peak. The sector can still be a good business. It is no longer the stock that pulls India.
The five-year return chart says the same thing in rupees. Three of the five leaders destroyed capital on price. The other two barely beat standing still. The revenue ranking did not change. The wealth-creation story did.
So: the five leaders were not dethroned as businesses. TCS is still the scale and margin leader. Infosys is still the other franchise. They were dethroned as automatic wealth creators.
What would change both charts is the same pair of numbers: growth that sits above inflation without a weaker rupee doing most of the work, and AI revenue that lifts the total rather than replacing the old book at a discount. Until those show up, treat the five as quality on sale, not as the automatic compounders they used to be. The right label for this page is not “India IT compounds.” It is “India IT resets.”
Educational commentary only. Not a recommendation to buy or sell any security.